Research And Implementation Of Brainwave Signal Visualization System

With the progress of brain science and technology,people learn more and more about brain activity.In recent years,people have been able to intuitively understand brain activity from EEG signals and magnetoencephalography(MEG).Among these signals,EEG signal has become a more widely used one because of its low cost and convenient use.On the other hand,with the breakthrough of artificial intelligence technology,new technologies including generative adversarial networks have been proposed and applied in various fields rapidly.Using new algorithms and models in artificial intelligence for brain wave signal analysis has become a promising field in the ascendant.Our research aims to visualize brainwave signals and explore human cognitive ability.The main contents of the article are as follows:In the processing of EEG,some traditional methods are introduced and changed to meet our needs.Then related technologies used in this paper are introduced,including brain-computer interface technology,data visualization technology,generation of adversarial network(GAN)and so on.Among them,the EEG signal acquisition in BCI technology and the BCI based on EEG signal are introduced.The application of EEG signal in epilepsy treatment is illustrated as an example.On the basis of data sets and related technologies,standard image classifier and EEG classifier are trained.The purpose is to judge the correctness of EEG generated image and lay a good foundation for EEG generated experiment.The EEG signal processing and brainwave visualization technology were studied.Many mathematical techniques such as correlation dimension,entropy and wavelet transform are introduced into traditional signal processing methods because EEG is a nonlinear and non-stationary stochastic process.It is worth mentioning that the traditional methods can also achieve real-time examination of brain diseases such as epilepsy and other medical applications.However,for the deep-seated human brain activities such as consciousness,the traditional way can not show the content of thinking in an intuitive way.In brainwave visualization,this paper introduces the rapidly developing generative adversarial networks GAN.First we collects certain specifications of EEG signals in appropriate ways,then we trains GAN with a certain loss function and C+D+G structure.C refers to the trained classifier,D refers to the discriminator trying to identify true and false pictures,and G refers to the generator that tries to generate false pictures with Gauss noise and deceives the discriminator network.It is worth mentioning that additional Gauss layer is adopted to solve the problem of too small sample set of EEG.Then we use the standard image classifier which has been trained to test the generated results,and the results show that the classification effect is good.